Hi all.
Am installing a different engine, not computer controlled in my car and the transmission torque converter wants 12 V only above 45MPH.
I think seeing somewhere an IC that compared its internal preset frequency and another fed to it; sort of a frequency comparator, not exactly a PLL.
There is an installed sensor that outputs 4 pulses per driveshaft revolution. I can make the calculations to find out how many Hertz that correspond to 45 MPH
Let's say 45MPH is 35 Hz. I want a "high" signal above 35Hz; "low" signal under 35Hz
How do I do it ? I strongly prefer to avoid µcontrollers.
EDN shows a 74HC4046 circuit; unsure if will do it ---> http://www.edn.com/article/CA265498.html
Miguel
How to sense over/under frequency ?
How to sense over/under frequency ?
- Abolish the deciBel ! -
Re: How to sense over/under frequency ?
Hi Miguel,
One thing about that circuit you linked to is that you may have to have inputs
that are 50 percent duty cycle.
Analog techniques arent as good as digital when measuring frequency.
This means a uC chip would be the best choice because you can get
good accuracy and only use one chip. But since you dont want to
go that route, another digital method would be to simply build a
frequency counter and instead of having the output go to a display
you would have it go to a digital comparator. You could set the
reference frequency with dip switches once you know what frequency
corresponds to your target speed. The comparator would output a
one "1" when it is above target and a zero "0" when below. That's
about as simple as it gets.
Also, you may not need to compare every bit of the result either,
depending on the set point accuracy desired.
Total chip count would be two to four 4 bit counters (or
up to two 8 bit counters), plus two 8 bit digital comparators.
That's around 4 to 6 ic chips...plus one or two 8 bit dip switches
unless you care to hardwire the set point in. Add an oscillator
and gate logic for the timing pulse.
One thing about that circuit you linked to is that you may have to have inputs
that are 50 percent duty cycle.
Analog techniques arent as good as digital when measuring frequency.
This means a uC chip would be the best choice because you can get
good accuracy and only use one chip. But since you dont want to
go that route, another digital method would be to simply build a
frequency counter and instead of having the output go to a display
you would have it go to a digital comparator. You could set the
reference frequency with dip switches once you know what frequency
corresponds to your target speed. The comparator would output a
one "1" when it is above target and a zero "0" when below. That's
about as simple as it gets.
Also, you may not need to compare every bit of the result either,
depending on the set point accuracy desired.
Total chip count would be two to four 4 bit counters (or
up to two 8 bit counters), plus two 8 bit digital comparators.
That's around 4 to 6 ic chips...plus one or two 8 bit dip switches
unless you care to hardwire the set point in. Add an oscillator
and gate logic for the timing pulse.
LEDs vs Bulbs, LEDs are winning.
Re: How to sense over/under frequency ?
I wouldn't think accuracy is important. I would consider saving the pulses in a capacitor that has a bleeder resistor and trip a detector if a certain voltage is reached.
Re: How to sense over/under frequency ?
i would say use a 8 pin PIC processor, and write a suiting program. would not be too hard, and very versatile as well
fripster
fripster
Once a WireHead, Always a WireHead
Re: How to sense over/under frequency ?
Hi,fripster wrote:i would say use a 8 pin PIC processor, and write a suiting program. would not be too hard, and very versatile as well
fripster
I thought the same, but he had already mentioned that he didnt really want to use a uC for this so
i offered another possible solution. I had hoped that this other solution also shows how much harder
it can be to accurately do a job that a uC can do in with chip.
LEDs vs Bulbs, LEDs are winning.
Re: How to sense over/under frequency ?
There is an old Ic which I think is still around, which can be used as a V to F or an F to V - now to remember it's part number, I think it was an LM something - used to be used a lot for rev counter projects.
Found it! http://www.jaycar.com.au/images_uploaded/LM2907.PDF
Colin
Found it! http://www.jaycar.com.au/images_uploaded/LM2907.PDF
Colin
On a clear disk you can seek forever.
Re: How to sense over/under frequency ?
I thought up something simple using a 4XXXB series CMOS D flip flop.
Have your incoming positive pulses charge up a capacitor through a diode. Connect a bleeder resistor to ground so that the capacitor voltage tends to bleed to zero with time.
Connect the capacitor voltage to the D input of a D flip flop.
Connect the incoming pulse signal to the clock input of the flip flop.
If the tach pulses are slow, the capacitor discharges to a point that the D input sees as logic zero when the next clock signal comes in. This zero state on the D input immediately before the clock pulse is transferred to the output Q terminal. The Q terminal will stay a constant zero.
If the tach pulses are fast, the capacitor still has a positive voltage that the D input sees as a logic one, and a high state is sent to the Q output. This is the opposite of the above paragraph and the Q output will remain at a constant logic high.
If you want to add a little hysteresis, add a high value resistor from the Q output to the D input.
There ya go! No mickey processors and no code to write. No memory to program. God, they'd put those things in scrub brushes if you let them.
Have your incoming positive pulses charge up a capacitor through a diode. Connect a bleeder resistor to ground so that the capacitor voltage tends to bleed to zero with time.
Connect the capacitor voltage to the D input of a D flip flop.
Connect the incoming pulse signal to the clock input of the flip flop.
If the tach pulses are slow, the capacitor discharges to a point that the D input sees as logic zero when the next clock signal comes in. This zero state on the D input immediately before the clock pulse is transferred to the output Q terminal. The Q terminal will stay a constant zero.
If the tach pulses are fast, the capacitor still has a positive voltage that the D input sees as a logic one, and a high state is sent to the Q output. This is the opposite of the above paragraph and the Q output will remain at a constant logic high.
If you want to add a little hysteresis, add a high value resistor from the Q output to the D input.
There ya go! No mickey processors and no code to write. No memory to program. God, they'd put those things in scrub brushes if you let them.
-=VA7KOR=- My solar system includes Pluto.
-
- Posts: 2277
- Joined: Wed Nov 24, 2004 1:01 am
- Location: ASHTABULA,OHIO
- Contact:
Re: How to sense over/under frequency ?
Have you considered a dual monostable such as CD4528/38. Reference (set pulse) and incoming RPM pulse both trigger each half of the chip. One shot times will be based on the range of incoming frequencys expected, but each of those one shot times will be exactly the same. I generally use a one shot pulse width of slightly less than the very highest incoming period expected. Use these outputs to charge up two different capacitors, each feeding a comparator's inputs.This circuit is quite accurate and have been used as frequency discriminators for quite a while. This will give you a high or low comparator output for tranny signal above or below set point.
Bodgys suggested LM2907 chip is a good one - but beware, they need close to a %50 duty cycle to perform properly. They don't tell you that in the data sheets. I had to learn the hard way
Bodgys suggested LM2907 chip is a good one - but beware, they need close to a %50 duty cycle to perform properly. They don't tell you that in the data sheets. I had to learn the hard way
Re: How to sense over/under frequency ?
I see the tendency to convert the pulses to a DC level and use a plain voltage comparator; or to use µcontrollers.
Bob; Robert, Colin :
What about this perhaps without the unnecessary ancillary parts of the circuit , uses the 4528 :
http://www.electronics-lab.com/projects ... index.html
I have to read it carefully and evaluate. Be back at some point... Thanks,
....engine went in today, will try to start it tomorrow...
Miguel
Bob; Robert, Colin :
What about this perhaps without the unnecessary ancillary parts of the circuit , uses the 4528 :
http://www.electronics-lab.com/projects ... index.html
I have to read it carefully and evaluate. Be back at some point... Thanks,
....engine went in today, will try to start it tomorrow...
Miguel
- Abolish the deciBel ! -
-
- Posts: 2277
- Joined: Wed Nov 24, 2004 1:01 am
- Location: ASHTABULA,OHIO
- Contact:
Re: How to sense over/under frequency ?
Yes, the CD4528, but preferably the CD 4538 - exact same function but has better stability and repeatability. Actually if your set point is a constant at 45MPH, and you aready have a built in pulse train available, You could determine what the voltage output will be from the 4528 at 45 MPH. Now just make your set point from a power supply derived DC voltage into one input of the comparator. The one shot output/DC coversion is fed into the other input to trip at 45 MPH.
Of coarse, I am assuming that the engine already has some sort of speed sensor already in place from your initial posting, so I am thinking that there is no need for another one. Sensor>4528>DC> comp input # 1. Comp input #2 is your predetermined DC out put that the above circuit will produce at 45 MPH. This will have to be determined by some experimentation after you have proto'ed the circuit, Want to change the 45 MPH setpoint up or down? - just dial in a different DC voltage to that comp's input.
Of coarse, I am assuming that the engine already has some sort of speed sensor already in place from your initial posting, so I am thinking that there is no need for another one. Sensor>4528>DC> comp input # 1. Comp input #2 is your predetermined DC out put that the above circuit will produce at 45 MPH. This will have to be determined by some experimentation after you have proto'ed the circuit, Want to change the 45 MPH setpoint up or down? - just dial in a different DC voltage to that comp's input.
Re: How to sense over/under frequency ?
Yes it does have some spare parts. It looks like the coil is an antenna that picks up and amplifies broadcasted ignition noise. You don't need the parts in the top half of the diagram.Externet wrote:What about this perhaps without the unnecessary ancillary parts of the circuit , uses the 4528 : http://www.electronics-lab.com/projects ... index.html
There are some different approaches you can take to design that speed detector.
"Whatever turns your crank". I have always strived to make mine as simple as possible.
Good luck with that. If it won't start right away, check the ignition firing order. GM, Ford, Chrysler number their cylinders differently. I've seen a couple of people frustrated before they checked that.Externet wrote:....engine went in today, will try to start it tomorrow...
Miguel
-=VA7KOR=- My solar system includes Pluto.
Re: How to sense over/under frequency ?
Hi all
Could a Missing Pulse Detector
Depending on the RC time of R2 & C1 (allows for trim) when the incomming frequency drops below frequency the 555 output will toggle low. The output could drive a small relay maybe?
Ron
Could a Missing Pulse Detector
Depending on the RC time of R2 & C1 (allows for trim) when the incomming frequency drops below frequency the 555 output will toggle low. The output could drive a small relay maybe?
Ron
Re: How to sense over/under frequency ?
Hi again,
If you dont mind the accuracy of analog, you can always set up a single one shot (like 555)
to output a pulse width of constant duration, triggered by the incoming signal, then
average the output of the one shot with a resistor and cap and compare that voltage
to a comparator who's other input is a pot to +Vcc and ground. You would then dial
in the correct set point with the pot.
Total parts would be the 555 plus a comparator chip plus a pot and some resistors and
a small capacitor. Under 5 dollars US.
If you dont mind the accuracy of analog, you can always set up a single one shot (like 555)
to output a pulse width of constant duration, triggered by the incoming signal, then
average the output of the one shot with a resistor and cap and compare that voltage
to a comparator who's other input is a pot to +Vcc and ground. You would then dial
in the correct set point with the pot.
Total parts would be the 555 plus a comparator chip plus a pot and some resistors and
a small capacitor. Under 5 dollars US.
LEDs vs Bulbs, LEDs are winning.
Re: How to sense over/under frequency ?
You might run your calculations again (using your actualExternet wrote:Let's say 45MPH is 35 Hz. I want a "high" signal above 35Hz; "low" signal under 35Hz
data for rim size and tyre size). 35Hz is very high for
normal highway speeds.
Here's a handy on-line calulator.
Where is the speed transducer (pick up)? On the drive shaft?
You'll need to know the rear diff ratio to scale the calculated
speed noted above.
Re: How to sense over/under frequency ?
Instead of measuring the frequency, measure the period.Externet wrote: Hi all.
Am installing a different engine, not computer controlled in my car and the transmission torque converter wants 12 V only above 45MPH.
...
Let's say 45MPH is 35 Hz. I want a "high" signal above 35Hz; "low" signal under 35Hz
For a non-uC solution take the pick up pulses and divide
them by two in a flip flop stage. These are now clean
50/50 pulses of duration proportional to speed.
Fire a non-retriggerable monostable from the output and
compare it's output to it's input. If both are the same
the vehicle is going above your threshold, if they are
not the same the vehicle is slower (or at rest).
You do not need any fancy PLL or other core analogue.
Who is online
Users browsing this forum: Ahrefs [Bot], Google [Bot] and 24 guests